To use inhomogeneous Magnetization Transfer (ihMT) to realize dipolar relaxation time (T_1D) weighted imaging, a new MR contrast.

The inclusion of a dipolar reservoir in the existing two pool model for MT allowed interpreting the inhomogeneous MT (ihMT) signal as a dipolar order effect - characterized by a relaxation time T_1D - within motion restricted molecules¹. Comparing saturation of dual frequency irradiation (MT^+- image), with the less efficient saturation of single frequency irradiation (MT⁺ image) allowed revealing dipolar order effect in tissues that have a nonzero T_1D. The large fraction of the semi-solid, bound magnetization with long T_1D values was responsible for the higher ihMT signal obtained in myelinated structures¹ and hence for the apparent specificity of ihMT for myelin^2,3. In this study, we demonstrate that an ihMT signal can actually be evidenced in any component with non-trivial T_1D value. Adjustment of the dual frequency irradiation efficiency by increase of Δt, the repetition rate of consecutive pulses, filters the signal of shorter T_1D components. This provides a means to realize T_1D-weighted imaging, a new source of MR contrast between tissues.IhMT experiments were performed at 11.75T (vertical MR system, Bruker, AV 500WB, transmit/receive volume coil: ∅ 2cm, length 3cm) on hair conditioner (containing lamellar liquid crystal structures⁴, LLC), agarose (agar 2% and 4%), and in vivo, on mouse (n=4, temperature 37±1°C). A 2D single-slice single-shot RARE sequence (TR/TE=3.4s/1.82ms, slice thickness=1mm, FOV=25x25 mm and mtx=64x64) combined with a pulsed ihMT preparation³ was used. Single frequency irradiation (Fig. 1a) was realized with a long train of Hann-shaped RF pulses (length, PW, frequency f=+8kHz for MT⁺ image, f=-8kHz for MT^- image), repeated every Δt for a total saturation time τ, overall depositing a total energy of saturation E_TR. Pulsed-dual frequency irradiation (MT^+- image) at identical E_TR value and achieved by alternating the frequency offset every other pulse from +8kHz to –8kHz (Fig. 1b) was used for LLC, agar and mouse experiments. Additionally, a strategy using cosine-modulated Hann-shape pulses for dual-frequency irradiation (Fig. 1c) was also tested in vivo. Values of saturation parameters are reported in table 1. IhMTR values, defined as the ihMT signal (MT⁺+MT^--2MT^+-) divided by the unsaturated M₀ image, were measured in LLC, agar and in brain internal capsule white matter (IC), cortical gray matter (GM) and muscle area (Mu). The contrast between IC and GM and between IC and Mu was evaluated by measuring the ratio of their ihMTR values. Different intensities of ihMTR and changes with Δt were obtained in LLC (T_1D^LLC~270ms, measured using the model developed in (1)) and agar2% (T_1D^agar2%~2.3ms) and 4% (T_1D^agar4%~2.6ms) (Fig. 2). High ihMTR values in LLC (ihMTR^LLC=31.5%, Δt=1.3ms) were only little affected by Δt increase (ihMTR^LLC=28.5%, Δt=6.3ms), whereas low agar ihMTR values obtained for Δt=1.3ms (ihMTR^agar2%=1.25%, ihMTR^agar4%=1.85%) were divided by almost two for Δt=3.3ms (ihMTR^agar2%=0.65%, ihMTR^agar4%=0.95%) and tended toward zero for Δt=5.3ms (ihMTR^agar2%<0.4%, ihMTR^agar4%<0.6%). In vivo pulsed dual-saturation results show an increase of ihMTR with decreasing Δt in all structures. For instance, for Δt decreasing from 3.3ms to 1.3ms, the relative ihMT signal increase was similar in IC and GM (~+20%) due to close T_1D values (T_1D^IC=5.9ms, T_1D^GM=5.2ms) and was greater in Mu (+55%, T_1D^Mu=2.1ms). This resulted in a constant IC/GM contrast but a loss of IC/Mu contrast with decreasing Δt (Fig. 3, left, plots and images). Compared to pulsed-dual saturation, the use of cosine-modulated pulses allowed quasi-instantaneous dual saturation. This led to significant and Δt-independent increase of ihMTR values in all structures: +22% in IC (ihMTR^IC~6% vs 4.8%), +30% in GM (ihMTR^GM~3.5% vs 2.7%) and +65% in Mu (ihMTR^Mu~3% vs 1.8%). This signal increase was however accompanied by important loss of IC/Mu contrast (-60%) (Fig. 3 right, plots and images), consistent with minimal T_1D-weighting.For pulsed-dual saturation, increasing Δt filters the ihMT signal of shorter T_1D components (e.g. agar) with little effect on the signal of long T_1D components (e.g. LLC). This filtering strategy allowed performing T_1D-weighted imaging, which can be used in vivo to enhance the specificity of ihMT and reinforce its contrast for longer T_1D tissue. Hence, ihMTR maps at Δt=3.3ms showed high specificity for WM (T_1D^Mu<Δt<T_1D^WM) compared to that at Δt=1.3ms (Δt <T_1D^Mu<T_1D^WM), however, at the cost of a slight decrease in sensitivity (ihMTR^IC=4.7% for Δt=1.3ms vs 4.0% for Δt=3.3ms). Conversely, short T_1D component (e.g. Mu) ihMT signal can be revealed using faster dual saturation approaches such as cosine-modulated pulses. This strategy could enable the study of ihMT in other tissues than brain.T_1D-weighted imaging, a new source of MR contrast between tissues, can be realized with ihMT.